( events)   Timezone:  
Oral
Mon Apr 30 03:15 PM -- 03:30 PM (PDT) @ Exhibition Hall A
Wasserstein Auto-Encoders
Ilya Tolstikhin · Olivier Bousquet · Sylvain Gelly · Bernhard Schoelkopf
[ PDF

We propose the Wasserstein Auto-Encoder (WAE)---a new algorithm for building a generative model of the data distribution. WAE minimizes a penalized form of the Wasserstein distance between the model distribution and the target distribution, which leads to a different regularizer than the one used by the Variational Auto-Encoder (VAE). This regularizer encourages the encoded training distribution to match the prior. We compare our algorithm with several other techniques and show that it is a generalization of adversarial auto-encoders (AAE). Our experiments show that WAE shares many of the properties of VAEs (stable training, encoder-decoder architecture, nice latent manifold structure) while generating samples of better quality.